The samples used in these studies were sourced and deidentified via Tissue Solutions Ltd., a subsidiary of BioIVT LLC which is accredited for the collection, storage, and commercial distribution of biospecimens by the Office for Human Research Protections (OHRP) of the United States Department of Health. The use of publicly available biospecimens from an accredited supplier is not considered to fall under research in human subjects, and as such, institutional review board (IRB) approval or exemption was not required. Prior to use, the toenails (n = 42), which had been frozen for storage, were thawed at room temperature for 10 min, immersed in 70% (v/v) ethanol for 1 min, and then washed twice by vortexing in deionized water. The four antifungal compounds tested were amorolfine (Loceryl® 5% lacquer), ciclopirox (Ciclopoli® 8% lacquer), naftifine (Exoderil® 1% solution), and tioconazole (Trosyd® 28% solution). Before treatment, the thickness of each toenail sample was measured with an Oditest caliper (Kroeplin GmbH, Schlüchtern, Germany). Five measurements were performed on different areas of each toenail and the mean of these measurements recorded.

For treatment, each toenail was laid on gauze moistened with sterile water in a petri dish. One antifungal was applied per toenail within a premarked zone of 1 cm2 on the nail surface at a dose of 10 µL/cm2. Nails were then allowed to air-dry for 30 min at 24 °C in a humid cell culture incubator, after which the surface of each toenail was cleaned with five cotton swabs wetted with acetone. The samples were then stored at −80 °C.

Nails stored at −80 °C were placed inside a cryostat maintained at −20 °C, and 10-µm nail sections were prepared for each sample. The midline cross sections of the toenails (region of interest) were mounted onto adhesive tape and placed on a MALDI plate. The mounted nail sections were then placed in a desiccator prior to matrix deposition to ensure the sections were dry. An optical image of the plate was acquired with a scanner to synchronize the positions of the nail sections with the laser target.

A stock solution of 10 mM amorolfine was prepared in 100% (v/v) dimethyl sulfoxide (DMSO), and stock solutions of 10 mM ciclopirox, naftifine, and tioconazole were prepared in 100% (v/v) ethanol. A dilution series of each antifungal was then prepared by diluting the stock solutions in a 1:1 (v:v) solution of ethanol and water; 1.0 μL of each dilution was spotted directly onto adhesive tape and placed in a desiccator for 15 min before MALDI matrix deposition. A spot of pure solvent was included as a negative control. The spotted calibration series was later placed onto the slides near the sections for image calibration.

For MALDI-FTCIR analysis, 2,5-dihydroxybenzoic acid matrix (DHB; 40 mg/mL in 1:1 methanol/water + 1% trifluoroacetic acid) was sprayed over the toenail sections with an automatic sprayer system (TM-Sprayer; HTX Technologies, Chapel Hill, NC, USA). The nail sections were then analyzed with a 7 T MALDI-FTICR instrument (SolariX; Bruker, Billerica, MA, USA) in the continuous accumulation of selected ions (CASI) positive mode centered on the targeted compound m/z at a spatial resolution of either 70 μm to cover the entire nail section or 200 μm for the dilution series to generate the calibration curves. All MSI acquisitions were performed with data reduction set between 0.30 and 0.50 (this was kept consistent among images of nails treated with the same antifungal), and one image was acquired per nail.

For the kinetic study of amorolfine penetration, the distribution profile of amorolfine was assessed 3, 6, 9, and 24 h after treatment in eight nails (n = 2 at each time point), and one control sample was used as a reference. For comparison of the amorolfine, ciclopirox, naftifine, and tioconazole distribution profiles after 3 h, 32 nails (n = 8 per antifungal agent) were evaluated, and one control sample was used as a reference.

The control nails were used to evaluate the tissue extinction coefficients (TECs) [12] of the antifungals for normalization. For this purpose, each antifungal was sprayed at a concentration of 5 μM on top of the control nail sections and next to the nail sections (directly onto the adhesive tape). Signals detected in the nail sections were compared with the signals beside the nail sections to determine the TECs.

The MALDI-FTICR data were analyzed with the flexImaging v.5.0 (Bruker), DataAnalysis v.5.0 (Bruker), and Multimaging v1.2 (Aliri, Loos, France) software packages. When evaluating antifungal distribution, the image intensity scale was adjusted to eliminate the background noise (by increasing the lower signal threshold) and enable optimal visualization (by decreasing the upper signal threshold). Convolution of the signal distributions was performed on the original images using a normalized uniform kernel that averaged the values around a position. The kernel size was manually optimized for the analysis to minimize background noise.

For absolute quantification of the antifungals within the nail sections, the MSI datasets obtained from the nail sections and dilution series were used. The TEC-based method in the Multimaging software was applied to these data to correct the signal in each region of interest and obtain the antifungal concentration in μg/g of tissue and the concentration in μM.

To quantitatively evaluate penetration of the antifungals through the nail samples, penetration profiles were generated for each treated nail section. The methodology was based on the following workflow: (1) the molecular distribution of the antifungal was first obtained by MSI as described above; (2) the region of interest for the penetration profile was selected (this was generally the entire region of acquisition, unless undesirable features present, such as contamination, poor histology, or folding of the sections, would significantly impact the results); (3) each pixel in the region of interest was re-aligned to define a common 0-μm position at the top of the nail section; (4) the mean concentration per raw pixel was calculated according to the quantitative MSI results and the depth was determined based on the spatial resolution for the acquisition (70 μm for each pixel); and (5) penetration profiles were then generated using the quantification and depth data.

Based on published data on the minimum inhibitory concentrations (MIC) of the four test compounds needed to inhibit 50% and 90% (MIC50 and MIC90) of Trichophyton rubrum, the fold differences between the MIC and the antifungal concentrations in the nails (termed the multiplicity of MIC) were calculated for each. The published MIC50 and MIC90 values (in µM) used for these calculations were 0.79 and 12.6 for amorolfine [13], 9.65 and 77.19 for ciclopirox [13], 0.39 and 12.35 for naftifine [13], and 1.29 and 2.58 for tioconazole [14], respectively.

The Shapiro–Wilk test was used to check the assumption of normality of data within groups. The Kruskal–Wallis H test was used to test for statistically significant differences across groups. When testing for statistically significant differences between groups two by two, the Student t test was used where data were normally distributed, and the Mann–Whitney U test was used where data were not normally distributed. A p value of < 0.05 was considered statistically significant. Python v3.8 (Python Software Foundation, Beaverton, OR, USA) was used for all statistical analyses.